def get_sw_scores():
res = {}
with open('data/scores.sw') as f:
for line in f:
line = line.strip()
if line.startswith('query:'):
cur_prot = line[6:]
res[cur_prot] = {}
elif line.startswith('score: '):
it = line[7:].split(' -- ')
res[cur_prot][it[1]] = int(it[0])
p = list(res.keys())
s = list()
for i in range(len(p)):
scores = np.zeros((len(p), ), dtype=np.float32)
for j in range(len(p)):
norm = max(res[p[i]][p[i]], res[p[j]][p[j]])
scores[j] = res[p[i]][p[j]] / norm
s.append(scores)
df = pd.DataFrame({'proteins': p, 'scores': s})
prots, sequences = read_fasta(open('data/swissprot.fasta', 'r'))
prots_dict = {}
for prot, seq in zip(prots, sequences):
prots_dict[prot] = seq
sequences = list()
for prot in p:
sequences.append(prots_dict[prot])
df['sequences'] = sequences
df.to_pickle('data/sw_scores.pkl')
return res
def pick_long_reads(files, length, outputfile, args):
totalcount = 0
passedcount = 0
for fastafile in files:
for header, seq in read_fasta(fastafile, False):
totalcount = totalcount + 1
if len(seq) >= length:
longread = True
if args.cut_stars:
seq = seq.split('*')
tmp_length = 0
for i in range(0, len(seq)):
if len(seq[i]) > tmp_length:
tmp_length = len(seq[i])
j = i
if tmp_length < length:
print header, tmp_length
longread = False
seq = seq[j]
if longread:
passedcount = passedcount + 1
if header.startswith("Contig"):
filename = (fastafile.split("/")[-1]).split(".")[0]
header = filename + "_" + header
outputfile.write('>%s\n%s\n' % (header, seq))
outputfile.close()
print 'searched %s sequences.\n \
%s of them were longer than %s' \
%(totalcount,passedcount,length)
def build_index(args):
i_file = args.input
utils.logging("[INFO] Start downloading reference file.", args)
tempbase = utils.gen_file()
utils.mkdir(tempbase)
reffile = os.path.join(tempbase, "raw.fa")
utils.read_hdfs(i_file, reffile)
tempfiles = [
open(os.path.join(tempbase, "%s.fa" % m), 'w') for m in conv_way
]
utils.logging("[INFO] Start transforming reference file.", args)
# read ref
for chrid, seq in utils.read_fasta(reffile):
for i, method in enumerate(conv_way):
(strand, a_from, a_to) = (method[0], method[2], method[4])
if strand == "W":
tempfiles[i].write(
">%s\n%s\n" %
(chrid,
seq.translate(utils.make_trans_with(strand, a_from,
a_to))))
else:
tempfiles[i].write(
">%s\n%s\n" %
(chrid,
seq.translate(utils.make_trans_with(strand, a_from,
a_to))[::-1]))
# close all files
for i, method in enumerate(conv_way):
tempfiles[i].close()
utils.logging("[INFO] Start launching bowtie2-build.", args)
# run bowtie jobs
procs = []
utils.mkdir(os.path.join(tempbase, "index"))
for i, method in enumerate(conv_way):
out_pref = os.path.join(tempbase, "index", method)
build_log = out_pref + ".build.log"
proc = Process(target=call_bowtie,
args=(
tempfiles[i].name,
out_pref,
build_log,
))
procs.append(proc)
proc.start()
for proc in procs:
proc.join()
utils.logging("[INFO] Start uploading index file.", args)
# move to hdfs
utils.copy_to_hdfs(tempbase, args.output, remove_original=True)
def align(sc, args):
import utils as g_utils
import align_utils as a_utils
## broadcast raw reference
ref_file = os.path.join(args.tempbase, "ref.fa")
g_utils.read_hdfs(os.path.join(args.ref, "raw.fa"), ref_file)
ref_dict = {}
for chrid, seq in g_utils.read_fasta(ref_file):
ref_dict[chrid] = (seq, len(seq))
g_utils.logging("[DEBUG] loading reference done", args)
bc_refdict = sc.broadcast(ref_dict)
## read from hadoop
readRDD = sc.textFile( args.input ) \
.map( lambda x: g_utils.line2kv( x))
if args.testmode == "balancing":
readRDD = readRDD.partitionBy(args.nodes)
readRDD = readRDD.cache()
## transform and get result of bowtie
c2tTransRDD = readRDD.mapValues(lambda x: (x[0].translate(
g_utils.make_trans_with("W", "C", "T")), x[1]))
c2tMapRDD = c2tTransRDD.mapPartitionsWithIndex(
lambda i, ptn: a_utils.mapping(i, "C2T", ["W_C2T", "C_C2T"], ptn, args
))
g2aTransRDD = readRDD.mapValues(lambda x: (x[0].translate(
g_utils.make_trans_with("W", "G", "A")), x[1]))
g2aMapRDD = g2aTransRDD.mapPartitionsWithIndex(
lambda i, ptn: a_utils.mapping(i, "G2A", ["W_G2A", "C_G2A"], ptn, args
))
mergedRDD = sc.union([readRDD, c2tMapRDD, g2aMapRDD])
combRDD = mergedRDD.combineByKey( lambda v: [v],\
lambda lst, v: lst + [v],\
lambda l1, l2: l1 + l2 )
filteredRDD = combRDD.mapValues( lambda x: a_utils.select_and_find_uniq_alignment( x))\
.filter( lambda (k, v): v is not None )
# .filter( lambda (k, v): not (v is None))
if args.testmode == "balancing":
filteredRDD = filteredRDD.partitionBy(args.nodes)
methylRDD = filteredRDD.map( lambda x: a_utils.calc_methyl(x, bc_refdict.value, args.num_mm) )\
.filter( lambda x: x is not None )
result_path = os.path.join(args.output, "alignment")
methylRDD.map(lambda x: a_utils.res_to_string(x)).saveAsTextFile(
result_path)
return result_path
def retrieve_orfs(orfs,fastaFile,orfFile):
orfFile = open(orfFile,'w')
for header,seq in read_fasta(fastaFile):
header = header.split()[0]
if header in orfs:
if orfs[header][2] == '+':
orfFile.write('>%s\n%s\n' %(header,seq[orfs[header][0]-1:orfs[header][1]]))
else:
rev_seq = reverse_complement(seq[orfs[header][0]-1:orfs[header][1]])
orfFile.write('>%s\n%s\n' %(header,rev_seq))
def tally_gc_counts(fasta_dict=read_fasta()):
results_dict = dict()
for k in sorted(fasta_dict.keys()):
counter = collections.Counter(fasta_dict[k])
total = sum(counter.values())
gc_count = counter['G'] + counter['C']
gc_percent = 100.0 * gc_count / total
results_dict[k] = gc_percent
return sorted(results_dict.items(), key=results_dict.get)[-1]
def run(args):
fasta = read_fasta(args.fasta, args.logfile)
gtf = read_gtf(args.gtf, args.logfile)
out("Random Indexing Test:", args.logfile)
for i, (ID, sequence) in enumerate(fasta.items()):
print(ID)
if "|" in ID:
ID = ID.split("|")[3]
gtf_items = gtf[ID]
if not gtf_items:
ID = ".".join(ID.split(".")[:-1])
gtf_items = gtf[ID]
print(ID)
gtf_items = gtf[ID]
exons = [(start, end) for _, start, end, _ in gtf_items]
print(gtf_items)
strand = gtf_items[0][3]
chrom = gtf_items[0][0]
out("Transcript ID: {}".format(ID), args.logfile)
out(
"len: {:4d} sequence[:20]: {}".format(len(sequence),
sequence[:20]), args.logfile)
out("gtf_items: {}".format(gtf_items), args.logfile)
# DEBUG: Checking that the exon lengths sum up to the length of the sequence
total_length = 0
for _, start, stop, _ in gtf_items:
assert stop > start
total_length += stop - start
assert total_length == len(sequence), "{} != {}\n{}\n".format(
total_length, len(sequence), ID, gtf_items)
assert strand == "+" or strand == "-", "{}".format(strand)
out("exons: {}".format(exons), args.logfile)
for _ in range(10):
queried_index = np.random.randint(0, len(sequence))
genomic_index = find_location(queried_index, exons, strand,
sequence)
out(
"Queried Index: {}\tGenomic index: {}".format(
queried_index, genomic_index), args.logfile)
out(
"Transcript window around queried index: {}".format(
sequence[queried_index - 10:queried_index + 11]),
args.logfile)
out("", args.logfile)
if i > 20:
exit()
out("", args.logfile)
def train(args):
'''
This is where the model is created & trained.
'''
# Loading the protein fasta sequences
orig_seqs = []
with open(args.input_file, 'r') as infasta:
for _, seq in utils.read_fasta(infasta):
orig_seqs.append(seq)
# Prepare training data
data_x, data_y, vocab_size, vocab_decode = \
utils.load_data(orig_seqs, args.seq_length)
# Creating the RNN-LSTM model
rnn_model = Sequential()
# Add first layer with input shape provided
rnn_model.add(
LSTM(args.hidden_dim,
input_shape=(None, vocab_size),
return_sequences=True))
# Add the remaining layers
for i in range(args.num_layers - 1):
rnn_model.add(LSTM(args.hidden_dim, return_sequences=True))
# We treat each char in the vocbulary as an independet time step
# and use TimeDistributed wrapper to apply the same dense layer (with
# number of units = vocab_size) and same weights to output one
# time step from the sequence for each time step in the input.
# In other words, we will process all the time steps of the input
# sequence at the same time.
rnn_model.add(TimeDistributed(Dense(vocab_size)))
rnn_model.add(Activation('softmax'))
rnn_model.compile(loss="categorical_crossentropy", optimizer="rmsprop")
# Train the model
rnn_model.fit(data_x,
data_y,
batch_size=args.batch_size,
verbose=1,
epochs=args.train_epochs)
# Generate new sequences
proteins = []
for i in range(args.generate_epochs):
new_pep = utils.generate_seq(rnn_model, args.seq_length, vocab_size,
vocab_decode)
proteins.append(new_pep)
# Write new protein sequences to a fasta file
utils.write_fasta(proteins, args.output_file)
def load_fasta_sequences(fasta_file1, fasta_file2, fasta1_to_fasta2_blast):
fasta_sequences = {
''.join(os.path.split(filename)[-1].split('.')[:-1]):
[[headerStr, seq] for headerStr, seq in read_fasta(filename)]
for filename in [fasta_file1, fasta_file2]
}
blast_results = load_blast_results(fasta1_to_fasta2_blast, [0, 1, 2])
fasta_sources = list(fasta_sequences.keys())
output_name = '-'.join(
['targets', f'{fasta_sources[1].split("_")[-1]}', 'blast'])
cols = [v[0] for v in fasta_sequences[fasta_sources[0]]]
rows = [v[0] for v in fasta_sequences[fasta_sources[1]]]
values = pd.DataFrame(0.0, index=rows, columns=cols)
for val in blast_results:
values.at[val[1], val[0]] = val[2]
values.to_csv(output_name + '.csv')
def main():
parser = argparse.ArgumentParser(
description='Adopt pairwise aligner to generate identity matrix')
parser.add_argument('fasta_path',
type=str,
help='the path of fasta files to parse')
parser.add_argument('out_path', type=str, help='output dir')
args = parser.parse_args([
'/home/ZwZ/database/M/linsi/M1_M2_unique.fasta',
'/home/ZwZ/database/M/linsi/M1_M2_unique.npy'
])
fasta = read_fasta(args.fasta_path)
identity_M = identity(fasta)
np.save(args.out_path, identity_M)
def bx_multiplicity(rfile):
"""
Calculate barcode multiplicity given a read file name. Returns a dictionary
of barcode -> num reads with that barcode.
"""
bxs = {}
if rfile == "-":
rfile = "/dev/stdin"
with open(rfile) as reads:
if not reads.isatty():
for _, _, bx, _ in read_fasta(reads):
if bx != None:
bxs.setdefault(bx, 0)
bxs[bx] += 1
else:
raise RuntimeError(
"Reads must be piped from stdin if file name is not provided")
return bxs
def insert_peptide_sequences_into_db(self, peptide_fasta):
"""
Insert peptide sequences into DB using peptide FASTA file.
NOTE: Risk for high memory consumption as entire FASTA is kept in memory.
"""
sequences = {
header.split()[0]: sequence
for header, sequence in read_fasta(peptide_fasta,
keep_formatting=False)
}
discriminative_peptides = self.db.execute(
"SELECT peptide FROM peptides").fetchall()
for peptide in discriminative_peptides:
self.db.execute(
"UPDATE peptides SET sequence = ? WHERE peptide = ?",
(sequences[peptide[0]], peptide[0]))
#logging.debug("Added sequence %s for peptide %s", sequences[peptide[0]], peptide[0]) # TODO: verbose
self.db.commit()
def main():
parser = argparse.ArgumentParser(
'filter out identical sequence and write a clean fasta file')
parser.add_argument(
'in_path',
type=str,
help='please give an absolute path of in put fasta file',
default='/home/ZwZ/database/M/linsi/M1_M2_all.fasta')
parser.add_argument('out_path',
type=str,
help='the absolute path of filtered fasta file',
default='/home/ZwZ/database/M/linsi/filter.fasta')
args = parser.parse_args()
ALL = utils.read_fasta(args.in_path, True)
keep, discard, _ = search_identical(ALL) # the most important func
all = [ALL[i] for i in keep]
SeqIO.write(all, args.out_path, 'fasta')
def load_data():
ngram_df = pd.read_pickle('data/ngrams.pkl')
vocab = {}
for key, gram in enumerate(ngram_df['ngrams']):
vocab[gram] = key + 1
gram_len = len(ngram_df['ngrams'][0])
print('Gram length:', gram_len)
print('Vocabulary size:', len(vocab))
ngrams = list()
proteins = list()
f = open('data/swissprot.fasta')
prots, seqs = read_fasta(f.readlines())
for protein, seq in zip(prots, seqs):
if not is_ok(seq) or len(seq) - gram_len + 1 > MAXLEN:
continue
proteins.append(protein)
grams = list()
for i in range(len(seq) - gram_len + 1):
grams.append(vocab[seq[i: (i + gram_len)]])
ngrams.append(grams)
df = pd.DataFrame({
'proteins': proteins,
'ngrams': ngrams,
})
def get_values(df):
grows = []
gcols = []
gdata = []
for i, row in enumerate(df.itertuples()):
for j in range(len(row.ngrams)):
grows.append(i)
gcols.append(j)
gdata.append(row.ngrams[j])
data = sparse.csr_matrix((gdata, (grows, gcols)), shape=(len(df), MAXLEN))
return data
return proteins, get_values(df)
def main():
min = 200
max = 370
#parameters that used
mis = 0
#read primer
fpri, rpri = utils.read_trad_primer(sys.argv[1])
lfile = sys.argv[2]
num = 0
seq = utils.read_fasta(sys.argv[2])
for item in seq.items():
se = item[1]
for f in fpri:
for i in range(0, len(se)-len(f)):
str1 = se[i:len(f)+i]
#print str1, f
#print mismatch(str1,f,mis)
if mismatch(str1,f,mis):
#print "match"
num += 1
output = ">F:"+lfile+"_"+str(num)+"\n"+f+'\n'
#print output
#print ">F:",f, i
for r in rpri:
rp = r
for j in range(0, len(se)-len(rp)):
str2 = se[j:len(rp)+j]
if mismatch(str2,rp,mis):
#print "match"
frp = reverse_complement.get_rc(rp)
#print ">R:",frp,j+len(rp), j+len(rp)-i
flen = j+len(rp)-i
#print flen
if(flen >min and flen<max):
output = output + ">R:"+lfile+"_"+str(num)+"\n"+frp+'\n'
print output
def _read(self, f):
return utils.read_fasta(f)
parser.add_argument("--step", "-s", default=15, help="the length between each two fragments")
parser.add_argument("--flen", "-fl", default=53, help="the length for each fragment")
parser.add_argument("--prob", "-p", default=0.5, help="the probability for selecting a target site")
parser.add_argument("--numsite", "-ns", default=0, help="the number of target sites")
args = parser.parse_args()
fasta_name = args.input1
onemirna = args.input2
outf = args.output
step = args.step
fraglen = args.flen
prob = args.prob
nums = args.numsite
headers, mrnaseqs = read_fasta(fasta_name)
# prepare the input data
mirnaf = open(onemirna, 'r')
mirnaseq = mirnaf.read().rstrip('\r\n').upper()
print(mirnaseq)
remirna = "".join(reversed(mirnaseq))
mirnaseq = remirna
print(remirna)
maxlen = 79
maxlenmir = 26
mirnalen = len(mirnaseq)
if mirnalen < maxlenmir:
import numpy as np
import random
#read in sites
posfile = '/Users/student/Documents/Algorithms/Alg_final_project/data/rap1-lieb-positives.txt'
negfile = '/Users/student/Documents/Algorithms/Alg_final_project/data/yeast-upstream-1k-negative.fa'
testfile = '/Users/student/Documents/Algorithms/Alg_final_project/data/rap1-lieb-test.txt'
poslist = util.read_pos(posfile)
finaltestlist = util.read_pos(testfile)
posreversecomp = []
for i in poslist:
posreversecomp.append(util.reverse_complement(i))
poslist = poslist + posreversecomp
neglist = util.read_fasta(negfile)
negreversecomp = []
for i in neglist:
negreversecomp.append(util.reverse_complement(i))
neglist = neglist + negreversecomp
for i in neglist:
if i in set(poslist):
neglist.remove(i)
#print('negs',neglist[:10])
print('neg', len(neglist))
print('pos', len(poslist))
shortneg = []
for i in neglist:
#TODO adapt so this is a random slice of the negative
def _read(self, f):
s, t = utils.read_fasta(f, dna_only=True)
return s, t
def load_sequences(self) -> list:
self.initial_alignment_df = read_fasta(self.file_name)
return [sequence for sequence in self.initial_alignment_df.get("sequence")]
parser.add_argument("--numsite",
"-ns",
default=0,
help="the number of called target sites")
args = parser.parse_args()
fasta_mrna = args.input1
fasta_mirna = args.input2
outf = args.output
step = args.step
fraglen = args.flen
prob = args.prob
nums = args.numsite
hmrna, mrnaseqs = read_fasta(fasta_mrna)
hmirna, mirnaseqs = read_fasta(fasta_mirna)
# prepare the input data
remirnas = []
for one in mirnaseqs:
oneremirna = "".join(reversed(one))
remirnas.append(oneremirna.upper())
print(mirnaseqs[0])
print(remirnas[0])
mirnaseqs = remirnas
maxlen = 79
maxlenmir = 26
mirnalen = len(min(mirnaseqs, key=len))
import numpy as np
from utils import read_fasta, estimate_population, generate_population
initial_alignment_df = read_fasta(filename='dataset/BB11001-m.fa')
# initial_alignment = [np.fromstring(sequence, dtype=np.uint8) for sequence in initial_alignment_df.get("sequence")]
initial_alignment = [
sequence for sequence in initial_alignment_df.get("sequence")
]
# initial_alignment = np.array(initial_alignment)
population = generate_population(size=20, alignment=initial_alignment)
print(population)
print(estimate_population(population))
#
# print(f'Estimated sequence value: {estimated_value}')
def main(args):
# Mapping COG id to proteins
cog_prot = defaultdict(lambda: defaultdict(list))
# Mapping protein to COG ids
prot_cog = defaultdict(lambda: defaultdict(list))
if not os.path.exists(args.dest):
print >> sys.stderr, 'Creating directory %s' % args.dest
os.makedirs(args.dest)
if not os.path.isdir(args.dest): sys.exit('Destination is not a directory')
if args.cog_csv:
cog_csv = (l.strip('\n').split(',') for l in args.cog_csv)
else:
if not os.path.exists('%s/cog2003-2014.csv' % args.dest):
print >> sys.stderr, 'Downloading COG csv from NCBI'
response = urllib2.urlopen(
'ftp://ftp.ncbi.nih.gov/pub/COG/COG2014/data/cog2003-2014.csv')
with open('%s/cog2003-2014.csv' % args.dest, 'w') as outh:
outh.write(response.read())
cog_csv = (l.strip('\n').split(',')
for l in open('%s/cog2003-2014.csv' % args.dest, 'r'))
for l in cog_csv:
protid = l[2]
cogid = l[6]
spos = int(l[4])
epos = int(l[5])
cog_prot[cogid][protid] = merge_interval_list(cog_prot[cogid][protid] +
[(spos, epos)],
dist=1)
prot_cog[protid][cogid] = merge_interval_list(prot_cog[protid][cogid] +
[(spos, epos)],
dist=1)
print >> sys.stderr, "Found {:,} COGs".format(len(cog_prot))
print >> sys.stderr, "Found {:,} protein ids".format(len(prot_cog))
c = Counter(len(v) for k, v in prot_cog.iteritems())
print >> sys.stderr, "Proteins belonging to multiple COGs:"
print >> sys.stderr, 'Num COGs | Count'
for k in range(10):
print >> sys.stderr, '%s%d' % (str(k + 1).ljust(12), c[k + 1])
print >> sys.stderr, '%s%d' % ('11+'.ljust(12),
sum(v for k, v in c.iteritems() if k > 10))
if args.fasta:
seqiter = (
(seqname, seq)
for seqname, seq in read_fasta(gzip.GzipFile(fileobj=args.fasta)))
else:
if not os.path.exists('%s/prot2003-2014.fa.gz' % args.dest):
print >> sys.stderr, 'Downloading protein sequences from NCBI'
response = urllib2.urlopen(
'ftp://ftp.ncbi.nih.gov/pub/COG/COG2014/data/prot2003-2014.fa.gz'
)
with open('%s/prot2003-2014.fa.gz' % args.dest, 'wb') as outh:
outh.write(response.read())
seqiter = ((seqname, seq) for seqname, seq in read_fasta(
gzip.open('%s/prot2003-2014.fa.gz' % args.dest, 'rb')))
# Clear out destination
purge_msg = True
for cogid in cog_prot.keys():
if os.path.exists('%s/%s/%s.faa' % (args.dest, cogid[:5], cogid)):
if purge_msg:
print >> sys.stderr, 'Purging existing sequence files'
purge_msg = False
os.remove('%s/%s/%s.faa' % (args.dest, cogid[:5], cogid))
# Create the COG files
numseqs = 0
filenames = {}
seqcounts = Counter()
for seqname, seq in seqiter:
try:
m = re.search('gi\|(\d+)\|ref', seqname)
protid = m.group(1)
except (AttributeError, ValueError):
print >> sys.stderr, 'Error parsing sequence name: "%s"' % seqname
continue
numseqs += 1
if not numseqs % 100000:
print >> sys.stderr, 'Processed %d proteins...' % numseqs
for cogid, ivs in prot_cog[protid].iteritems():
for istart, iend in ivs:
newseqname = 'cog|%s|%s (%d-%d)' % (cogid, seqname, istart,
iend)
newseq = seq[istart - 1:iend]
seqcounts[cogid] += 1
if not os.path.exists('%s/%s' % (args.dest, cogid[:5])):
os.makedirs('%s/%s' % (args.dest, cogid[:5]))
with open('%s/%s/%s.faa' % (args.dest, cogid[:5], cogid),
'a') as outh:
print >> outh, '>%s\n%s' % (newseqname, fmt_seq(newseq))
filenames[cogid] = '%s/%s.faa' % (cogid[:5], cogid)
print >> sys.stderr, "Processed %d total proteins" % numseqs
with open('%s/fungroups.txt' % args.dest, 'w') as outh:
for k in sorted(filenames.keys()):
print >> outh, '%s\t%s\t%d' % (k, filenames[k], seqcounts[k])
the label of the next string.
In Rosalind's implementation, a string in FASTA format will be labeled by the ID "Rosalind_xxxx", where
"xxxx" denotes a four-digit code between 0000 and 9999.
Given: At most 10 DNA strings in FASTA format (of length at most 1 kbp each).
Return: The ID of the string having the highest GC-content, followed by the GC-content of that string. Rosalind
allows for a default error of 0.001 in all decimal answers unless otherwise stated; please see the note on
absolute error below.
"""
from utils import read_fasta, gc_content
if __name__ == '__main__':
sequences = read_fasta('data/rosalind_gc.txt')
max_gc = 0
max_id = None
for id, s in sequences.items():
gc = gc_content(s)
if gc > max_gc:
max_gc = gc
max_id = id
print max_id
print max_gc * 100
downstream_model_names = [
'56_Linear.pth', '66_Linear.pth', '68_Linear.pth', '71_Linear.pth',
'69_Linear.pth'
]
downstream_models = [
torch.load(os.path.join(model_path, model_name), map_location=device)
for model_name in downstream_model_names
]
pdb_file = 'C:\\Workspace\\PE\\UI\\saved_models\\1emm_mod.pdb'
atom_lines = utils.process_pdb(pdb_file, atoms_type=['N', 'CA', 'C', 'O'])
coord_array_ca, acid_array, coord_array = utils.extract_coord(
atom_lines, atoms_type=['N', 'CA', 'C', 'O'])
utils.compare_len(coord_array, acid_array, ['N', 'CA', 'C', 'O'])
seq_dict = utils.read_fasta(seq_file)
with open(os.path.join(files_path, 'results.txt'), 'a') as f:
f.write('sequence_name\tpredicted_value\n')
for seq_name, seq in seq_dict.items():
seq_dict[seq_name] = utils.seq2array(seq)
array = utils.get_knn_135(coord_array, seq)
input_ = torch.tensor(array.reshape(-1, 135)).float()
hidden = pretrained_model(input_).squeeze(1)
output = 0
for model in downstream_models:
model.eval()
model.is_training = False
output += model(hidden)
def write_fasta(self):
self.initial_alignment_df = read_fasta(self.file_name)
with open("BB11001.aln", "w") as file_handler:
for index, id in enumerate(self.initial_alignment_df.get("id")):
file_handler.write(">" + id + "\n")
file_handler.write(self.best_solution[index] + "\n")
def parse_orfs(orfFinderOut,orfFile,minLength):
orfOut = open(orfFile,'w')
for header, seq in read_fasta(orfFinderOut,False):
if len(seq) > minLength:
seq = find_common_startcodon(seq,minLength)
orfOut.write('>%s\n%s\n' %(header,seq))
parser.add_argument('--depth', default=7, help="min,max")
parser.add_argument('--period', default=1, help="min, max")
parser.add_argument('-k', default=1, help="k-fold validation")
parser.add_argument('model', nargs='+')
args = parser.parse_args()
seqs = []
labels = dict()
all_labels = set()
for f in args.model:
label = ".".join(os.path.split(f)[-1].split(".")[:-1])
all_labels.add(label)
for seq in utils.read_fasta(f):
seqs.append(seq)
labels[seq["seqid"]] = label
best_args = [12, 7, 1, 100]
print "width\tdepth\tperiod\terror_rate"
for w in get_range(args.window):
for d in get_range(args.depth):
if w < d:
continue
for p in get_range(args.period):
errors = utils.cross_validate(seqs, labels, int(args.k), window=w, depth=d, period=p)
avg_error = sum(errors) / len(errors)
def main(go_file, train_sequences_file, train_annotations_file,
test_sequences_file, test_annotations_file, out_terms_file,
train_data_file, test_data_file, min_count):
logging.info('Loading GO')
go = Ontology(go_file, with_rels=True)
logging.info('Loading training annotations')
train_annots = {}
with open(train_annotations_file, 'r') as f:
for line in f:
it = line.strip().split('\t')
prot_id = it[0]
if prot_id not in train_annots:
train_annots[prot_id] = set()
go_id = it[1]
train_annots[prot_id].add(go_id)
logging.info('Loading training sequences')
info, seqs = read_fasta(train_sequences_file)
proteins = []
sequences = []
annotations = []
for prot_info, sequence in zip(info, seqs):
prot_id = prot_info.split()[0]
if prot_id in train_annots:
proteins.append(prot_id)
sequences.append(sequence)
annotations.append(train_annots[prot_id])
prop_annotations = []
cnt = Counter()
for annots in annotations:
# Propagate annotations
annots_set = set()
for go_id in annots:
annots_set |= go.get_anchestors(go_id)
prop_annotations.append(annots_set)
for go_id in annots_set:
cnt[go_id] += 1
df = pd.DataFrame({
'proteins': proteins,
'sequences': sequences,
'annotations': prop_annotations,
})
logging.info(f'Train proteins: {len(df)}')
logging.info(f'Saving training data to {train_data_file}')
df.to_pickle(train_data_file)
# Filter terms with annotations more than min_count
res = {}
for key, val in cnt.items():
if val >= min_count:
ont = key.split(':')[0]
if ont not in res:
res[ont] = []
res[ont].append(key)
terms = []
for key, val in res.items():
terms += val
logging.info(f'Number of terms {len(terms)}')
logging.info(f'Saving terms to {out_terms_file}')
df = pd.DataFrame({'terms': terms})
df.to_pickle(out_terms_file)
logging.info('Loading testing annotations')
test_annots = {}
with open(test_annotations_file, 'r') as f:
for line in f:
it = line.strip().split('\t')
prot_id = it[0]
if prot_id not in test_annots:
test_annots[prot_id] = set()
go_id = it[1]
test_annots[prot_id].add(go_id)
logging.info('Loading testing sequences')
info, seqs = read_fasta(test_sequences_file)
proteins = []
sequences = []
annotations = []
for prot_info, sequence in zip(info, seqs):
prot_id = prot_info.split()[0]
if prot_id in test_annots:
proteins.append(prot_id)
sequences.append(sequence)
annotations.append(test_annots[prot_id])
prop_annotations = []
for annots in annotations:
# Propagate annotations
annots_set = set()
for go_id in annots:
annots_set |= go.get_anchestors(go_id)
prop_annotations.append(annots_set)
df = pd.DataFrame({
'proteins': proteins,
'sequences': sequences,
'annotations': prop_annotations,
})
logging.info(f'Test proteins {len(df)}')
logging.info(f'Saving testing data to {test_data_file}')
df.to_pickle(test_data_file)
assert len(set(len(x) for x in fasta.values())) <= 1
profile = collections.defaultdict(lambda: [0] * seq_len)
for i in range(seq_len):
for base in 'ACGT':
for dna_seq in fasta.values():
if dna_seq[i] == base:
profile[base][i] += 1
return dict(profile)
def calculate_consensus():
seq_len = len(list(fasta.values())[0])
consensus = []
profile = build_profile_matrix()
for j in range(seq_len):
max_count = 0, ''
for base in profile:
if profile[base][j] > max_count[0]:
max_count = profile[base][j], base
consensus.append(max_count)
return ''.join([i[1] for i in consensus])
fasta = read_fasta()
print(calculate_consensus())
for base, profile in build_profile_matrix().items():
print('%s: %s' % (base, ' '.join(str(i) for i in profile)))
def _read(self, f):
dnas = [list(dna) for _, dna in utils.read_fasta(f)]
return np.asarray(dnas)
from sys import argv, exit
from utils import read_fasta as read_fasta
if len(argv) < 2:
print("usage: script.py FASTA discriminative_peptides.txt")
exit(1)
fastafile = argv[1]
discpeps = argv[2]
disc = {}
with open(discpeps) as f:
for line in f:
if line.startswith("-"):
break
peptide, families = line.strip().split(" ", maxsplit=1)
translation = str.maketrans("{'}", " ")
families = families.translate(translation).split(",")
disc[peptide] = [fam.strip() for fam in families]
disc_seqs = {}
for header, sequence in read_fasta(fastafile):
try:
disc_seqs[sequence] = ", ".join(disc[header])
except KeyError:
pass
for seq, target in disc_seqs.items():
print(seq, target, sep="\t")
def run(args):
sequences = read_fasta(args.fasta,
args.logfile) if args.fasta is not None else None
if args.gtf is not None:
ref = read_gtf(args.gtf, args.logfile)
elif args.bed is not None:
ref = read_bed(args.bed, args.logfile)
if ref is None:
raise NotImplementedError
meRanCall = read_meRanCall(args.input, args.logfile)
outfile = open(args.output, "w")
errfile = open(args.error_out, "w")
header = "# <chromosome> <genomic position> <strand> <methylation rate> <transcript ID> <transcript position>\n"
outfile.write(header)
sequence = None
not_found_errors = 0
not_found_error_transcripts = set()
sequence_length_check_errors = 0
sequence_length_check_error_transcripts = set()
for ID, refPos, refStrand, methRate in meRanCall:
if sequences is not None:
sequence = sequences[ID]
out("ID:{} position:{}".format(ID, refPos), args.logfile)
if "|" in ID:
ID = ID.split("|")[3]
if ID not in ref:
out("Error: Could not find ID {} in reference file".format(ID),
args.logfile)
errfile.write(
"Error: Could not find ID {} in reference file\n".format(ID))
not_found_errors += 1
not_found_error_transcripts.add(ID)
continue
coordinates = ref[ID]
assert refStrand == "+"
if sequence is not None:
total_length = 0
for _, start, stop, _ in coordinates:
total_length += stop - start
if total_length != len(sequence):
out(
"Error: The lengths of the exons do not match the overall length of the sequence: exons {} != seq {}"
.format(total_length, len(sequence)), args.logfile)
if len(sequence) < total_length:
errfile.write(
"Error: The lengths of the exons of ID {} do not match the overall length of the sequence: exons {} != seq {}\n"
.format(ID, total_length, len(sequence)))
sequence_length_check_errors += 1
sequence_length_check_error_transcripts.add(ID)
continue
out("Attempting to handle Poly-A...", args.logfile)
length_diff = len(sequence) - total_length
last_chunk = sequence[-length_diff:]
if last_chunk == "A" * length_diff:
sequence = sequence[:-length_diff]
out("Handled this case by cutting off Poly-A",
args.logfile)
assert total_length == len(sequence)
else:
out("Could not handle this case by cutting off Poly-A",
args.logfile)
errfile.write(
"Error: The lengths of the exons of ID {} do not match the overall length of the sequence: exons {} != seq {}\n"
.format(ID, total_length, len(sequence)))
sequence_length_check_errors += 1
sequence_length_check_error_transcripts.add(ID)
continue
strand = coordinates[0][3]
chrom = coordinates[0][0]
genomic_position = find_location(refPos, coordinates)
line = "\t".join(
(chrom, str(genomic_position), strand, str(methRate), ID,
str(refPos))) + "\n"
outfile.write(line)
out("Finished writing to: {}".format(args.output), args.logfile)
out(
"Number of sites whose transcript was not found in the reference file: {}"
.format(not_found_errors), args.logfile)
out(
"Number of sites whose transcript sequence length did not equal the sum of all the exons: {}"
.format(sequence_length_check_errors), args.logfile)
out("Total meRanCall lines: {}".format(len(meRanCall)), args.logfile)
out("", args.logfile)
out(
"Number of transcripts that were not found in the reference file: {}".
format(len(not_found_error_transcripts)), args.logfile)
out(
"Number of transcripts whose length did not equal the sum of all the exons: {}"
.format(len(sequence_length_check_error_transcripts)), args.logfile)
out(
"Total number of transcripts read in the reference file: {}".format(
len(ref)), args.logfile)
errfile.write(
"Number of sites whose transcript was not found in the reference file: {}\n"
.format(not_found_errors))
errfile.write(
"Number of sites whose transcript sequence length did not equal the sum of all the exons: {}\n"
.format(sequence_length_check_errors))
errfile.write("Total meRanCall lines: {}\n".format(len(meRanCall)))
errfile.write("\n")
errfile.write(
"Number of transcripts that were not found in the reference file: {}\n"
.format(len(not_found_error_transcripts)))
errfile.write(
"Number of transcripts whose length did not equal the sum of all the exons: {}\n"
.format(len(sequence_length_check_error_transcripts)))
errfile.write(
"Total number of transcripts read in the reference file: {}\n".format(
len(ref)))
def _read(self, f):
return utils.read_fasta(f, dna_only=True)[0]
letter, _ = c.most_common(1)[0]
consensus += letter
return consensus, profile
def display(symbols, consensus, profile):
s = consensus + "\n"
for symb, counts in zip(symbols, profile):
s = s + f"{symb}: {list_to_string(counts)}\n"
return s
if __name__ == "__main__":
symbols = "ACGT"
dnas = [
"ATCCAGCT",
"GGGCAACT",
"ATGGATCT",
"AAGCAACC",
"TTGGAACT",
"ATGCCATT",
"ATGGCACT",
]
display(symbols, *profile(symbols, dnas))
dnas = read_fasta("inps/rosalind_cons.txt")
print(dnas)
s = display(symbols, *profile(symbols, dnas))
with open("outs/rosalind_cons.txt", "w") as f:
f.write(s)
